How much VRAM does 14B need?

Q4_0 is exactly 8,176,548,512 bytes (7.62 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

How large is 14B's KV cache?

25.00 GiB at 32K context with an f16 cache, computed per layer from the model's architecture. Quantizing the cache to q8_0 roughly halves it.

CausalLM · text

14B

CausalLM/14B

14B at Q4_0 is exactly 8,176,548,512 bytes (7.62 GiB / 8.18 GB) — an effective 4.617 bits per weight, not the nominal 4. Its KV cache at 32K is 25.00 GiB.

From the file· summed from 1 file(s)From the file· KV per layer

Parameters

14.2B

Architecture

llama

40 layers

Context

8,192

native (config.json)

License

wtfpl

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
Q4_0	7.62 GiB	8,176,548,512	4.617	—	TheBloke
Q4_1	8.39 GiB	9,013,279,392	5.090	—	TheBloke
Q5_0	9.17 GiB	9,850,010,272	5.562	—	TheBloke
Q5_1	9.95 GiB	10,686,741,152	6.035	—	TheBloke
Q8_0	14.02 GiB	15,058,954,912	8.504	—	TheBloke

KV cache by context

computed per layer

Context	KV cache (f16)●	Flat formula	Overstated by	Full / windowed / recurrent
4,096	3.13 GiB	3.13 GiB	—	40 / 0 / 0
8,192	6.25 GiB	6.25 GiB	—	40 / 0 / 0
16,384	12.50 GiB	12.50 GiB	—	40 / 0 / 0
32,768	25.00 GiB	25.00 GiB	—	40 / 0 / 0
65,536	50.00 GiB	50.00 GiB	—	40 / 0 / 0
131,072	100.00 GiB	100.00 GiB	—	40 / 0 / 0

Compare with

same modality, comparable size

vs Qwen3-Coder-30B-A3B-Instruct vs Qwen3.6-27B vs Qwen3.8-27B vs gemma-4-12B-it-qat-q4_0-unquantized vs gemma-4-E4B-it vs Qwen3-30B-A3B-Thinking-2507 All alternatives →Will it fit on my card? →

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_0 at roughly 7.42 GiB. The real file is 7.62 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json

Layers

Attention heads

KV heads

Head dim

128

Hidden size

5120

Vocab

152,064

Sliding window

none

SWA period

—

MLA

Experts

—

Experts per token

—

use_sliding_window

—

Questions people ask

How much VRAM does 14B need?: Q4_0 is exactly 8,176,548,512 bytes (7.62 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is 14B's KV cache?: 25.00 GiB at 32K context with an f16 cache, computed per layer. Quantizing the cache to q8_0 roughly halves it, which is often the difference between a context length fitting and not.
Which quantization of 14B should I use?: Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.